Thromboelastography and prediction of venous thromboembolism in orthopedic inpatients: development and internal validation of a clinical prediction model
This study developed and internally validated a clinical prediction model for venous thromboembolism in orthopedic inpatients, finding that while thromboelastography (TEG) parameters reflect a hypercoagulable phenotype, they do not significantly improve risk prediction beyond a parsimonious model based on age, length of stay, sex, and fracture/trauma status.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
In the world of medicine, the body's ability to stop bleeding is a delicate balance, a system that must be precise enough to seal a wound instantly yet flexible enough to keep blood flowing freely through the veins. When this balance tips too far toward clotting, a dangerous condition known as venous thromboembolism can occur. This is a situation where a blood clot forms in a deep vein, often in the leg, and can break loose to travel to the lungs, causing a potentially fatal blockage. For patients admitted to orthopedic wards for broken bones or joint surgeries, the risk of developing these clots is particularly high due to the trauma of injury, the stress of surgery, and the immobility that follows. Doctors have long relied on standard patient information, such as age and how long a person has been in the hospital, to guess who might be at risk. However, a newer technology called thromboelastography has emerged as a potential tool to see the clotting process in action, offering a dynamic picture of how fast blood clots form and how strong they become. The question facing modern medicine is whether this complex, real-time view of blood clotting offers a significant advantage over the simple, everyday facts doctors already know.
A team of researchers at Shiyan Taihe Hospital in China set out to answer this question by looking at the records of 1,186 consecutive patients admitted to their orthopedic department between January and October 2025. They wanted to see if adding the detailed data from thromboelastography to a standard patient profile would help them predict who would develop a blood clot more accurately than the profile alone. The researchers gathered a wide range of information for each patient, including their age, sex, the reason for their admission, and how long they stayed in the hospital. Crucially, they also collected six specific measurements from the thromboelastography tests performed within two days of admission. These measurements tracked different stages of the clotting process, from the initial time it took for a clot to start forming, to the speed at which it grew, the final strength of the clot, and how quickly the body began to break it down.
The study found that among the 1,186 patients, 64 developed a confirmed blood clot during their hospital stay, a rate of about 5.4 percent. When the researchers looked at the data without any complex adjustments, they saw a clear pattern: patients who developed clots generally had blood that clotted faster and formed stronger clots compared to those who did not. Specifically, the time it took for a clot to start was shorter, the clot grew more quickly, and the final structure was denser. This confirmed that the technology was indeed picking up on a biological state of hyper-coagulability, where the blood is primed to clot. However, the real test came when the researchers tried to build a prediction model. They first created a simple model using only the basic clinical facts: the patient's age, the length of their hospital stay, their sex, and whether they had a fracture or trauma. This simple model was already quite good at distinguishing between those who would and would not develop a clot.
When the team added the four most relevant thromboelastography measurements to this simple clinical model, the improvement in prediction accuracy was so small it was statistically indistinguishable from random chance. The ability of the model to correctly identify high-risk patients increased by a tiny fraction, a change so minor that it did not meet the threshold for being considered a meaningful improvement. In fact, the researchers found that once they knew a patient's age and how long they were staying in the hospital, the extra information from the blood test added essentially no new value to the prediction. The simple model, relying only on age and length of stay, performed just as well as the complex model that included the advanced blood testing. The researchers also checked if this result held true for different types of patients, such as those with fractures versus those with other conditions, and found that the performance remained consistent across the board.
The study concludes that while the advanced blood test successfully identifies the biological signs of a body that is prone to clotting, it does not help doctors predict who will actually suffer a clotting event any better than the basic information they already have. The researchers suggest that for the general population of orthopedic patients, routinely ordering these expensive and time-consuming tests to screen for risk may not be a cost-effective use of medical resources. Instead, the most efficient approach appears to be focusing on the straightforward clinical factors of age and hospital stay duration. The authors emphasize that their findings are based on a single hospital and a retrospective look at past records, meaning that further studies involving multiple centers and prospective designs are needed before these conclusions can be applied universally. Until such broader validation occurs, the routine addition of this sophisticated blood testing for risk stratification in unselected orthopedic patients remains an open question, with the current evidence pointing toward the sufficiency of simpler, readily available data.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.